Documentation
Everything you need to install, configure, and get the most out of SnapShield AI.
SnapShield AI is a Windows desktop application for annotating and redacting screenshots before you share them. It combines a 13-tool annotation editor, on-device AI redaction, batch processing, and optional Google Drive sync into a single lightweight binary.
All AI processing runs locally on your CPU — no internet required and no data ever leaves your machine unless you explicitly upload to your own Google Drive.
Version 1.6.4 for Windows 10 (1903+) and Windows 11, 64-bit. Both builds are identical; only the way they are installed differs.
Program Files
asks for administrator rights, because replacing the executable requires them.
Lite: no activation needed. Pro: up to 5 devices per key.
Captures the entire virtual screen — spanning all connected monitors — in one shot. Default hotkey: PrintScreen.
Captures the currently focused application window, trimmed to its exact client boundaries. Default hotkey: Alt+PrintScreen.
Draws a semi-transparent overlay over your screen. Drag to define a rectangle; release to capture exactly that area. Hotkey: Ctrl+PrintScreen.
Automatically scrolls the target window and stitches frames together using pixel-overlap detection. Works with browsers, PDF viewers, Office apps, and most Win32 applications. Hotkey: Ctrl+Shift+S.
Uses the Chrome DevTools Protocol to capture the complete scrollable content of a Chromium browser tab (Chrome, Edge, Brave) as a single pixel-perfect PNG — no manual scrolling, no animation artifacts. Works with the active tab automatically.
Press Ctrl+V with SnapShield in focus to paste any image from your clipboard directly into the editor.
SnapShield's AI stack runs entirely on your CPU — no internet connection is required and no image data leaves your machine. It uses two engines working in parallel:
Minimum 4 GB RAM recommended for AI features. Processing time depends on image size and CPU speed.
If detection reports “nothing sensitive found”
Windows OCR needs a language pack for the language in your screenshot. If the relevant one is not installed, the text pass finds nothing and SnapShield reports a clean image — which looks identical to a genuinely clean one. If you get that result on a screenshot you know contains an email address or an account number, open Settings → Time & language → Language & region in Windows, and check that your language has the optional Optical character recognition feature installed. Face detection is unaffected: it uses the bundled ONNX model, not OCR.
The UltraFace ONNX model locates every face in the image. SnapShield then places a blur annotation over each detected region automatically. Results are non-destructive — you can resize, delete, or adjust any annotation before exporting.
The Windows OCR engine maps every visible text block in the image and marks each one for redaction. Works on screenshots of UIs, code, terminal output, documents, and photos of printed text.
After OCR extracts text, SnapShield applies regex pattern matching to find and auto-redact:
Run face detection and text detection simultaneously in one pass. Results are merged and deduplicated — overlapping annotations are collapsed into a single region. This is the fastest way to fully sanitize a screenshot before sharing.
A lighter alternative to full auto-redact. Smart Blur obscures only the detected sensitive regions, leaving the rest of the image crisp and legible. Useful for screenshots where context around sensitive data still matters.
Choose the visual style applied to each redacted region:
Process an entire folder of screenshots with one command. Add files via drag-and-drop or the file picker, choose your detection mode (faces, text, PII, or all three), and SnapShield redacts every image in parallel. A live progress bar shows completion; the job can be cancelled at any time without data loss.
After annotating your screenshot, open the Export dialog and choose Upload to Google Drive (instead of saving to disk). SnapShield uploads the PNG to your Drive and returns a shareable link immediately.
You can also upload from the Cloud Sync panel (View → Cloud Sync) by dropping any image file onto it.
The Cloud Sync panel lists all images you have uploaded through SnapShield, with the file name, upload date, file size, and a one-click Copy Link button. You can also delete any uploaded file directly from Drive using the delete button — the file is removed from your Google Drive permanently.
Go to Settings → Cloud Sync → Disconnect. This clears the locally stored tokens and removes the account link from SnapShield. Your existing files in Google Drive are not affected. You can reconnect the same or a different account at any time.
Lossless. Lite exports include a small "SnapShield AI" text watermark; Pro does not.
Configurable quality (1–100, default 90). Lite includes watermark.
Modern format with excellent quality-to-size ratio.
Single-page PDF v1.4 embedding the annotated image at full resolution.
Ctrl+C copies the annotated image to the Windows clipboard for instant pasting.
Verify you are running Windows 10 version 1903 or later (64-bit). SnapShield AI is not yet Authenticode signed, so Windows SmartScreen shows "Windows protected your PC" on first run — click More info → Run anyway. If it still will not start, open the log from More ▸ Open Log Folder, or find it directly at %LocalAppData%\SnapShieldAI\snapshield.log, and send it to support.
Check Settings → Hotkeys to confirm the shortcut is assigned. Another application may be consuming the same key combination — try reassigning to an unused hotkey. For elevated processes (e.g. Task Manager), SnapShield must also run as administrator to capture hotkeys.
Reduce the scroll speed in Settings → Capture → Scrolling. Ensure the page is fully loaded before starting. Some browser extensions inject overlays that disrupt scroll detection — try disabling them. For web pages, use Full-Page Web Capture (Pro) instead, which uses CDP and never loses content.
Confirm you are on the Pro tier (AI Auto-Redact is not available on Lite). Open Settings → AI → Verify model files to check that face_detection.onnx is present. Ensure at least 4 GB of RAM is available. Very small or very low-contrast text may not be detected by OCR.
Check that the destination folder exists and that your user account has write permissions to it. Try exporting as PNG first. If exporting to PDF or WebP, ensure you are on Pro. If the problem persists, update to the latest version via Help → Check for Updates.
Copy and paste the key from your purchase email — do not retype it manually. Ensure you are entering the key in Settings → License → Activate, not in a search box. If you have reached the device limit for your tier, deactivate an unused device first.
Can't find what you're looking for?
Email us at info@nexoradia.com — we respond within one business day.
Go to Support →